International Inter-observer Variability of Breast Density Assessment

Abstract

Objective: To determine variability in visually assessed mammographic breast density categorization among radiologists practicing in Indonesia, the Netherlands, South Africa, and the United States.

Methods: Two-hundred consecutive 2D full field digital screening mammograms performed from September-December 2017 were selected and retrospectively reviewed from four global locations for a total of 800 mammograms. Three breast radiologists in each location (team) provided consensus density assessments of all 800 mammograms using BI-RADS® density categorization. Inter-reader agreement was compared using Gwet's AC2 with quadratic weighting across all four density categories and Gwet's AC1 for binary comparison of combined not dense versus dense categories. Variability of distribution between teams was calculated using the Stuart-Maxwell test of marginal homogeneity across all four categories and using McNemar's test for not dense versus dense categories. To compare readers from a particular country on their own 200 mammograms versus the other three teams, density distribution was calculated using conditional logistic regression.

Results: For all 800 mammograms, inter-reader weighted agreement for distribution among four density categories is 0.86 (Gwet's AC2 with quadratic weighting 95%, CI 0.85-0.88) and for not dense vs dense categories it is 0.66 (Gwet's AC1 95%, CI 0.63-0.70). Density distribution across four density categories was significantly different when teams were compared to each other and one team versus the other three teams combined (p<0.001). Overall, all readers placed the largest number of mammograms in the scattered and heterogeneous categories.

Conclusion: While reader teams from four different global locations had almost perfect inter-reader agreement in BI-RADS® density categorization, variability in density distribution across four categories remains statistically significant.